Discovery Inventories

Designs land better when the platform knows what your organization already runs. Discovery inventories the existing estate: APIs, datastores, cloud resources, organizational structure, knowledge stores, and file stores, so generation can reason about reuse and drift detection can compare the approved design against reality.

Plan Availability

Discovery is available on the Large plan and above. Connections are configured by an organization admin using read-only credentials.

Metadata Only, by Design

Discovery's rule is structural: it records what exists and its shape, never the contents.

• Datastores: engines, schemas, and structures; never rows.
• File stores: the tree and object metadata; file contents are never read.
• Organizational structure: aggregate team shapes and role counts; never person-level records.
• APIs and cloud resources: interface definitions and resource inventories.

This is what lets discovery inform architecture without becoming a data-governance problem of its own.

What It Powers

Consumer	Effect
Generation	The design can prefer what already runs (reuse over reinvention) and mark technologies as in use
Environment drift	Approved components with no discovered counterpart, and unsanctioned technology running outside any approved design, raise findings
Delivery drift	"Done in the tracker but not observed in the environment" needs discovery's view to fire

Discovery Makes Drift Honest

Without discovery, drift can only compare the design against itself and your tracker. With it, the comparison reaches the estate: the difference between "we approved a queue" and "a queue is actually running".